Attempts to elicit active immune responses in cancer patients to date can be classified as "non-specific" (i.e., the use of BCG) or "specific", i.e., the use of tumor cells, tumor cell extracts, mixtures of antigens from cell culture supernatant fluids or oncolysates of tumor cells. The vast majority of these efforts have been pursued in patients with metastatic melanoma. The development of a recombinant vaccine implies the use of specific and defined gene products or epitopes of an immunogen or an immunostimulatory molecule. Recombinant vaccines can also be used for "gene therapy" in that the latter approach requires using cells from a given patient and the insertion of a gene for an immunostimulatory molecule such as B7.1, B7.2, IL-2 or GM-CSF into those cells, either in situ or for administration of cultured cells back to the patient.
Recombinant vaccines can take many forms. Recombinant proteins can be synthesized by vectors such as baculovirus (an insect vector) or in eukaryotic cells. Synthetic peptides can also serve as immunogens. Peptide vaccines which consist of 9 to several dozen amino acids can take two forms. They can be mixed with adjuvant or they can be used to pulse peripheral blood cells as antigen presenting cells (APCs) for reinfusion into the patient. Recombinant vaccines can also be constructed by inserting the gene which codes for a given tumor associated antigen into a vector. Some of the common vectors used are vaccinia virus, avian pox viruses such as fowlpox or canary pox, BCG, adenovirus and Salmonella. These vectors, each with their advantages and disadvantages are usually employed because of the immunogenicity of their constitutive proteins, thus rendering the protein or epitope of the inserted gene more immunogenic. Recombinant vaccines can also take the form of an anti-idiotype antibody which is directed against a monoclonal antibody prepared against a given tumor associated antigen. Most recently, polynucleotide vaccines have been prepared which consist of naked DNA of a tumor associated gene in a plasmid containing a promoter gene. Whereas all of the above have been analyzed in animal models, very few studies have compared relative efficiencies of one approach versus the other. Clinical trials have now begun using some of these approaches in breast cancer and other carcinoma patients and others will most likely begin in the near future.
There are several antigens that have now been identified for potential use in recombinant vaccines for cancer therapies. The first of these is the c-erbB/2 oncogene which is found to be over expressed in approximately 20-30% of breast tumors (Pietras R J et al. Oncogene (9:1829-1838, 1994). It has been shown (Bernards R et al. Proc. Natl. Acad. Sci. USA 84:6854-6858, 1987) that the point mutated c-erbB/2 oncogene in rats, when inserted into vaccinia virus, is immunogenic and can lead to anti-tumor effects. The human c-erbB/2, however, is not mutated. It has recently been shown (Disis M L et al., Cancer Res. 54:1071-1076, 1994)), that this gene contains several epitopes which appear to generate human T-cell responses in vitro. The point mutated p53 oncogene, also found in many human breast tumors has been shown to be a potential target for cytotoxic T-cells (Yanuck M et al. Cancer Res. 53:3257-3261, 1993). Clinical studies are now beginning in which peptides reflecting specific point mutations are being pulsed with human peripheral blood lymphocytes (PBLs) and readministered to patients. The breast cancer mucin, MUC-1 or DF3, represents a differentiation antigen of the breast (Abe M and Kufe D, Proc. Natl. Acad. Sci. USA 90:282-286, 1993). While MUC-1 is expressed in a range of normal epithelial tissues, it appears to be uniquely glycosylated in breast cancer tissue. The tandem repeat of the core protein of the MUC-1 mucin has been reported to be immunogenic in humans (Barnd D L et al., Proc. Natl. Acad. Sci. USA 86:7159-7163, 1989) in that lymph nodes of breast cancer patients contain T-cells which can be activated by MUC-1 peptides in an non-MHC restricted manner. It has also been shown (Rughetti A et al., Cancer Res. 53:2457-2459, 1993) that ovarian cancer patients can make antibody responses to this region. Animal models in which the MUC-1 gene has been inserted into vaccinia virus have been reported (Hareuveni M et al., Proc. Natl. Acad. Sci. USA 87:9498-9502, 1990; Hareuveni et al., Vaccine 9:618-626, 1991). A clinical trial in which MUC-1 peptide is being pulsed with human PBLs is currently underway in breast cancer patients. Another mucin that represents a potential target for cancer therapy is TAG-72 which is found on approximately 70-80% of human breast cancers (Thor A et al., Cancer Res. 46:3118-3124, 1986).
Most attempts at active immunization against cancer antigens have involved whole tumor cells or tumor cell fragments, though it would be most desirable to immunize specifically against unique tumor antigens that distinguish malignant from normal cells. The molecular nature of the tumor associated antigens recognized by T lymphocytes is poorly understood. In contrast to antibodies that recognize epitopes or intact proteins, T cells recognize short peptide fragments (8-18 amino acids) that are present on cell surface class I or II major histocompatibility (MHC) molecules and it is likely that tumor associated antigens are presented and recognized by T cells in this fashion.
A number of genes have been identified that encode melanoma tumor antigens recognized by TIL in the context of the HLA-A2 class I molecule (Kawakami Y. et al. Proc. Natl. Acad. Sci. USA 91:3515-3519, 1994; Kawakami Y. et al. J. Exp. Med 180:347-352, 1994; Kawakami Y. et al. Cancer Res. 54:3124-3126, 1994).
The human carcinoembryonic antigen (CEA) also represents a potential target molecule for the immunotherapy of a range of human carcinomas including colorectal, gastric, pancreatic, breast, and non-small cell carcinomas (Robbins P F et al., Int. J. Cancer 53:892-897, 1993; Esteban J M et al., Cancer 74:1575-1583, 1994). Experimental studies have shown that anti-idiotype antibodies directed against anti-CEA monoclonal antibodies can elicit immune responses in mice (Bhattacharya-Chatterjee M et al., Int. Rev. Immuno. 7:289-302, 1991). Clinical studies using this anti-idiotype antibody are currently in progress. A recombinant vaccine has also been developed in which the CEA gene has been inserted into vaccinia virus (Kantor J. et al., J. Natl. cancer Inst. 84:1084-1091, 1992). A Phase I clinical trial involving this vaccine has just been completed.
The identification of an immunodominant peptide that represents a unique tumor antigen has opened new possibilities for immunization against cancer. Substantial evidence exists in animal models that immunization with immunodominant viral peptides can induce viral specific CTL that can confer protection against viral infection. Although pure peptide alone is ineffective in stimulating T cell responses, peptides emulsified in adjuvants or complexed with lipids have been shown to prime mice against challenge with fresh virus and can induce virus specific CTL that protect mice against lethal viral inocula (Kast, W. M. et al Proc. Nat'l Acad. Sci. U.S.A. 88:2283-2287, 1991; Deres, K. et al Nature 342:561-564, 1989; Gao, X. M. et al J. Immunol. 147:3268-3273, 1991; Aichele, P. J. Exp. Med. 171:1815-1820, 1990; Collins, D. S. et al J. Immunol. 148:3336-3341, 1992). Immunization of mice against splenocytes coated with Listeria monocytogenes peptide epitopes also results in the generation of Listeria specific CTL which can be expanded in culture. Adoptive transfer of these CTL can protect mice against lethal bacterial challenge (Harty, J. T. et al J. Exp. Med. 175:1531-1538, 1992). Peptides representing antigenic epitopes of HIV gp120 and gp160 emulsified in complete Freund's adjuvant can also prime specific CTL responses (Takahashi, H. et al Proc. Nat'l Acad. Sci. U.S.A. 85:3105-3109, 1988; Hart, M. K. et al Proc. Nat'l Acad. Sci. U.S.A. 88:9448-9452, 1991).
While immunization with peptides in adjuvants or complexed with lipids gives rise to T cell responses in mice, the reactions are rarely strong enough to induce T reactive cells in primary splenocytes. The detection of sensitized lymphocytes almost invariably requires secondary in vitro stimulation.
The expression of the B7 gene family has been shown to be an important mechanism of antitumor responses in both mice and humans. It is now becoming apparent that at least two signals are required for activation of naive T-cells by antigen bearing target cells: an antigen specific signal, delivered through the T-cell receptor, and an antigen independent or costimulatory signal leading to lymphokine products (Hellstrom, K. E. et al. Annals NY Acad. Sci. 690:225-230, 1993). Two important costimulatory molecules are B7-1, which is the ligand for T-cell surface antigens CD28 and CTLA4 (Schwartz, R. H. Cell 71:1065-1068, 1992; Chen, L. et al. Cell 71:1093-1102, 1992; Freeman, G. J. et al. J. Immunol 143:2714-2722, 1989; Freeman, G. J. et al. J. Exp. Med. 174:625-631, 1991), and B7-2, an alternative ligand for CTLA4 (Freeman, G. J. et al. Science 262:813-960, 1993). To date, both murine B7-1 and B7-2 (Freeman, G. J. et al. J. Exp. Med. 174:625-631, 1991; Freeman, G. J. et al. Science 262:813-960, 1995) and human B7-1 and B7-2 have been described (Freeman, G. J. et al. J. Immunol 143:2714-2722, 1989; Freeman, G. J. et al Science 262:909-911, 1993). It is unclear at this time whether the costimulatory signals provided by B7-1 and B7-2 are functionally distinct or redundant mechanisms for T-cell activation (Hathcock, K. S. et al. J. Exp. Med. 180:631-640, 1994). Most murine and human tumors do not express B7-1 or B7-2, implying that even when a tumor expresses a potential rejection antigen, it is unlikely to activate antitumor T-cell responses (Hellstrom, K. E. et al Annals. N.Y. Acad. Sci. 690:225-230, 1993); Hellstrom, I. Annals. N.Y. Acad. Sci. 690:24-31, 1993). In essence, anergy may result from only one signal being received by the T-cell (Hellstrom, K. e. et al. Annals. N.Y. Acad. Sci. 690:225-230, 1993. Transfection of B7 into melanoma cells was found to induce the rejection of a murine melanoma in vivo (Townsend, S. E. et al Science 259:368-370, 1993).
Vaccinia viruses have been extensively used in humans and the use of a vaccina based vaccine against smallpox has led to the worldwide eradication of this disease (reviewed in reference Moss, B. Science 252:1662-1667, 1991). Vaccinia viruses have the advantages of low cost, heat stability and a simple method of administration. Attempts have been made to develop vaccinia virus vectors for the prevention of other diseases.
Vaccina virus is a member of the pox virus family of cytoplasmic DNA viruses. DNA recombination occurs during replication of pox viruses and this has been used to insert DNA into the viral genome. Recombinant vaccina virus expression vectors have been extensively described. These vectors can confer cellular immunity against a variety of foreign gene products and can protect against infectious diseases in several animal models. Recombinant vaccina viruses have been used in human clinical trials as well. Cooney et al immunized 35 healthy HIV seronegative males with a recombinant vaccinia virus expressing the gp160 envelope gene of HIV (Cooney, E.. The Lancet 337:567-572, 1991). Graham et al randomized 36 volunteers to receive either recombinant vaccinia virus containing the gp160 HIV envelope protein or control vaccinia virus (Graham, B. S. et al J. Infect. Dis. 166:244-252, 1992). Phase I studies using recombinant vaccinia virus have begun in patients with metastatic melanoma using a recombinant virus expressing the p97 melanoma antigen (Estin, C. D. et al Proc. Nat'l Acad. Sci. 85:1052-1056, 1988) and a Phase I trial to use recombinant vaccinia virus expressing the human carcinoembryonic antigen in patients with advanced breast, lung or colorectal carcinoma has just been completed. In these trials, vaccinia virus is administered by intradermal scarification and side effects have been minimal including local skin irritation, lymphadenopathy and transient flu-like symptoms.
Fowlpox and canarypox viruses are members of the pox virus family (avipox virus genes). These viruses will only replicate in avian cells and cannot replicate in human cells. They are cytoplasmic viruses that do not integrate into the host genome but are capable of expression of a large number of recombinant genes in eukaryotic cells.
Recombinant avian pox virus expressing rabies glycoprotein has been used to protect mice, cats and dogs against live rabies virus challenge. Immunization of chickens and turkeys with a recombinant avian pox expressing the influenza HA antigen protected against a lethal challenge with influenza virus (Taylor et al., Vaccine 6:504-508, 1988). Canarypox volunteers received doses up to 10.sup.5.5 infectious units (Cadoz M., et al.., The Lancet 339: 1429-1432, 1992). In a recent trial sponsored by NIAID (Protocol 012A: A Phase I safety and immunogenicity trial of live recombinant canarypox-gp 160 MN (ALVAC VCP125 HIV-1gp160MNO in HIV-1 uninfected adults) patients received recombinant canarypox virus containing the HIV gp160 gene by intramuscular injection at doses up to 10.sup.5.5 pfu with little or no toxicity (personal communication, P. Fast, NIAID).
Avian pox viruses thus represent attractive vehicles for immunization since they can stimulate both humoral and cellular immunity, can be economically produced in high titer (10.sup.9 pfu/ml) and yet their inability to productively infect human cells substantially increases the safety of their use.
Another considerable advantage of avian pox virus is that there may be little or no cross-reactivity with vaccinia virus and thus previously vaccinated humans may not have preexisting immune reactivity to fowlpox virus proteins.